An embedded device for real-time noninvasive intracranial pressure estimation

Abstract

Monitoring of intracranial pressure (ICP) is key in many neurological conditions for diagnosis and guiding therapy. Current monitoring methods are highly invasive, limiting their use to the most critically ill patients. Based on a previously developed approach to noninvasive ICP estimation from cerebral blood flow velocity (CBFV) and arterial blood pressure (ABP) waveforms, I have implemented the algorithm on an embedded device (LPC4337 microcontroller) that can produce real-time estimates of ICP from noninvasively-obtained ABP and CBFV measurements. I have also fabricated a medical device prototype complete with peripheral interfaces for ABP and CBFV monitoring hardware and display and recording functionality for clinical use and post-acquisition analysis. The current device produces a mean estimate of ICP once per minute and can perform the necessary computations in 410 ms, on average. Real-time estimates of noninvasive ICP differed from the original batch-mode MATLAB implementation of the algorithm by 0.34 mmHg (RMSE). The contributions of this thesis take a step toward the goal of real-time noninvasive ICP estimation in a variety of clinical settings.